Toxicology Test Methods
The word biomarker is short for biological marker, a term that refers to a measurable event occurring in a biological system, such as the human body. This event is then interpreted as a reflection, or marker, of a more general state of the organism or of life expectancy. In occupational health, a biomarker is generally used as an indicator of health status or disease risk.
Biomarkers are used for in vitro as well as in vivo studies that may include humans. Usually, three specific types of biological markers are identified. Although a few biomarkers may be difficult to classify, usually they are separated into biomarkers of exposure, biomarkers of effect or biomarkers of susceptibility (see table 1).
Table 1. Examples of biomarkers of exposure or biomarkers of effect that are used in toxicological studies in occupational health
|Adipose tissue||Dioxin||Dioxin exposure|
|Exhaled breath||Toluene||Toluene exposure|
|Blood||Carboxyhaemoglobin||Carbon monoxide exposure|
|Red blood cells||Zinc-protoporphyrin||Lead exposure|
|White blood cells||DNA adducts||Mutagen exposure|
Given an acceptable degree of validity, biomarkers may be employed for several purposes. On an individual basis, a biomarker may be used to support or refute a diagnosis of a particular type of poisoning or other chemically-induced adverse effect. In a healthy subject, a biomarker may also reflect individual hypersusceptibility to specific chemical exposures and may therefore serve as a basis for risk prediction and counselling. In groups of exposed workers, some exposure biomarkers can be applied to assess the extent of compliance with pollution abatement regulations or the effectiveness of preventive efforts in general.
Biomarkers of Exposure
An exposure biomarker may be an exogenous compound (or a metabolite) within the body, an interactive product between the compound (or metabolite) and an endogenous component, or another event related to the exposure. Most commonly, biomarkers of exposures to stable compounds, such as metals, comprise measurements of the metal concentrations in appropriate samples, such as blood, serum or urine. With volatile chemicals, their concentration in exhaled breath (after inhalation of contamination-free air) may be assessed. If the compound is metabolized in the body, one or more metabolites may be chosen as a biomarker of the exposure; metabolites are often determined in urine samples.
Modern methods of analysis may allow separation of isomers or congeners of organic compounds, and determination of the speciation of metal compounds or isotopic ratios of certain elements. Sophisticated analyses allow determination of changes in the structure of DNA or other macromolecules caused by binding with reactive chemicals. Such advanced techniques will no doubt gain considerably in importance for applications in biomarker studies, and lower detection limits and better analytical validity are likely to make these biomarkers even more useful.
Particularly promising developments have occurred with biomarkers of exposure to mutagenic chemicals. These compounds are reactive and may form adducts with macromolecules, such as proteins or DNA. DNA adducts may be detected in white blood cells or tissue biopsies, and specific DNA fragments may be excreted in the urine. For example, exposure to ethylene oxide results in reactions with DNA bases, and, after excision of the damaged base, N-7-(2-hydroxyethyl)guanine will be eliminated in the urine. Some adducts may not refer directly to a particular exposure. For example, 8-hydroxy-2´-deoxyguanosine reflects oxidative damage to DNA, and this reaction may be triggered by several chemical compounds, most of which also induce lipid peroxidation.
Other macromolecules may also be changed by adduct formation or oxidation. Of special interest, such reactive compounds may generate haemoglobin adducts that can be determined as biomarkers of exposure to the compounds. The advantage is that ample amounts of haemoglobin can be obtained from a blood sample, and, given the four-month lifetime of red blood cells, the adducts formed with the amino acids of the protein will indicate the total exposure during this period.
Adducts may be determined by sensitive techniques such as high-performance lipid chromatography, and some immunological methods are also available. In general, the analytical methods are new, expensive and need further development and validation. Better sensitivity can be obtained by using the 32P post labelling assay, which is a nonspecific indication that DNA damage has taken place. All of these techniques are potentially useful for biological monitoring and have been applied in a growing number of studies. However, simpler and more sensitive analytical methods are needed. Given the limited specificity of some methods at low-level exposures, tobacco smoking or other factors may impact significantly on the measurement results, thus causing difficulties in interpretation.
Exposure to mutagenic compounds, or to compounds which are metabolized into mutagens, may also be determined by assessing the mutagenicity of the urine from an exposed individual. The urine sample is incubated with a strain of bacteria in which a specific point mutation is expressed in a way that can be easily measured. If mutagenic chemicals are present in the urine sample, then an increased rate of mutations will occur in the bacteria.
Exposure biomarkers must be evaluated with regard to temporal variation in exposure and the relation to different compartments. Thus, the time frame(s) represented by the biomarker, that is, the extent to which the biomarker measurement reflects past exposure(s) and/or accumulated body burden, must be determined from toxicokinetic data in order to interpret the result. In particular, the degree to which the biomarker indicates retention in specific target organs should be considered. Although blood samples are often used for biomarker studies, peripheral blood is generally not regarded as a compartment as such, although it acts as a transport medium between compartments. The degree to which the concentration in the blood reflects levels in different organs varies widely between different chemicals, and usually also depends upon the length of the exposure as well as time since exposure.
Sometimes this type of evidence is used to classify a biomarker as an indicator of (total) absorbed dose or an indicator of effective dose (i.e., the amount that has reached the target tissue). For example, exposure to a particular solvent may be evaluated from data on the actual concentration of the solvent in the blood at a particular time following the exposure. This measurement will reflect the amount of the solvent that has been absorbed into the body. Some of the absorbed amount will be exhaled due to the vapour pressure of the solvent. While circulating in the blood, the solvent will interact with various components of the body, and it will eventually become subject to breakdown by enzymes. The outcome of the metabolic processes can be assessed by determining specific mercapturic acids produced by conjugation with glutathione. The cumulative excretion of mercapturic acids may better reflect the effective dose than will the blood concentration.
Life events, such as reproduction and senescence, may affect the distribution of a chemical. The distribution of chemicals within the body is significantly affected by pregnancy, and many chemicals may pass the placental barrier, thus causing exposure of the foetus. Lactation may result in excretion of lipid-soluble chemicals, thus leading to a decreased retention in the mother along with an increased uptake by the infant. During weight loss or development of osteoporosis, stored chemicals may be released, which can then result in a renewed and protracted “endogenous” exposure of target organs. Other factors may affect individual absorption, metabolism, retention and distribution of chemical compounds, and some biomarkers of susceptibility are available (see below).
Biomarkers of Effect
A marker of effect may be an endogenous component, or a measure of the functional capacity, or some other indicator of the state or balance of the body or organ system, as affected by the exposure. Such effect markers are generally preclinical indicators of abnormalities.
These biomarkers may be specific or non-specific. The specific biomarkers are useful because they indicate a biological effect of a particular exposure, thus providing evidence that can potentially be used for preventive purposes. The non-specific biomarkers do not point to an individual cause of the effect, but they may reflect the total, integrated effect due to a mixed exposure. Both types of biomarkers may therefore be of considerable use in occupational health.
There is not a clear distinction between exposure biomarkers and effect biomarkers. For example, adduct formation could be said to reflect an effect rather than the exposure. However, effect biomarkers usually indicate changes in the functions of cells, tissues or the total body. Some researchers include gross changes, such as an increase in liver weight of exposed laboratory animals or decreased growth in children, as biomarkers of effect. For the purpose of occupational health, effect biomarkers should be restricted to those that indicate subclinical or reversible biochemical changes, such as inhibition of enzymes. The most frequently used effect biomarker is probably inhibition of cholinesterase caused by certain insecticides, that is, organophosphates and carbamates. In most cases, this effect is entirely reversible, and the enzyme inhibition reflects the total exposure to this particular group of insecticides.
Some exposures do not result in enzyme inhibition but rather in increased activity of an enzyme. This is the case with several enzymes that belong to the P450 family (see “Genetic determinants of toxic response”). They may be induced by exposures to certain solvents and polyaromatic hydrocarbons (PAHs). Since these enzymes are mainly expressed in tissues from which a biopsy may be difficult to obtain, the enzyme activity is determined indirectly in vivo by administering a compound that is metabolized by that particular enzyme, and then the breakdown product is measured in urine or plasma.
Other exposures may induce the synthesis of a protective protein in the body. The best example is probably metallothionein, which binds cadmium and promotes the excretion of this metal; cadmium exposure is one of the factors that result in increased expression of the metallothionein gene. Similar protective proteins may exist but have not yet been explored sufficiently to become accepted as biomarkers. Among the candidates for possible use as biomarkers are the so-called stress proteins, originally referred to as heat shock proteins. These proteins are generated by a range of different organisms in response to a variety of adverse exposures.
Oxidative damage may be assessed by determining the concentration of malondialdehyde in serum or the exhalation of ethane. Similarly, the urinary excretion of proteins with a small molecular weight, such as albumin, may be used as a biomarker of early kidney damage. Several parameters routinely used in clinical practice (for example, serum hormone or enzyme levels) may also be useful as biomarkers. However, many of these parameters may not be sufficiently sensitive to detect early impairment.
Another group of effect parameters relate to genotoxic effects (changes in the structure of chromosomes). Such effects may be detected by microscopy of white blood cells that undergo cell division. Serious damage to the chromosomes—chromosomal aberrations or formation of micronuclei—can be seen in a microscope. Damage may also be revealed by adding a dye to the cells during cell division. Exposure to a genotoxic agent can then be visualized as an increased exchange of the dye between the two chromatids of each chromosome (sister chromatid exchange). Chromosomal aberrations are related to an increased risk of developing cancer, but the significance of an increased rate of sister chromatid exchange is less clear.
More sophisticated assessment of genotoxicity is based on particular point mutations in somatic cells, that is, white blood cells or epithelial cells obtained from the oral mucosa. A mutation at a specific locus may make the cells capable of growing in a culture that contains a chemical that is otherwise toxic (such as 6-thioguanine). Alternatively, a specific gene product can be assessed (e.g., serum or tissue concentrations of oncoproteins encoded by particular oncogenes). Obviously, these mutations reflect the total genotoxic damage incurred and do not necessarily indicate anything about the causative exposure. These methods are not yet ready for practical use in occupational health, but rapid progress in this line of research would suggest that such methods will become available within a few years.
Biomarkers of Susceptibility
A marker of susceptibility, whether inherited or induced, is an indicator that the individual is particularly sensitive to the effect of a xenobiotic or to the effects of a group of such compounds. Most attention has been focused on genetic susceptibility, although other factors may be at least as important. Hypersusceptibility may be due to an inherited trait, the constitution of the individual, or environmental factors.
The ability to metabolize certain chemicals is variable and is genetically determined (see “Genetic determinants of toxic response”). Several relevant enzymes appear to be controlled by a single gene. For example, oxidation of foreign chemicals is mainly carried out be a family of enzymes belonging to the P450 family. Other enzymes make the metabolites more water soluble by conjugation (e.g., N-acetyltransferase and μ-glutathion-S-transferase). The activity of these enzymes is genetically controlled and varies considerably. As mentioned above, the activity can be determined by administering a small dose of a drug and then determining the amount of the metabolite in the urine. Some of the genes have now been characterized, and techniques are available to determine the genotype. Important studies suggest that a risk of developing certain cancer forms is related to the capability of metabolizing foreign compounds. Many questions still remain unanswered, thus at this time limiting the use of these potential susceptibility biomarkers in occupational health.
Other inherited traits, such as alpha1-antitrypsin deficiency or glucose-6-phosphate dehydrogenase deficiency, also result in deficient defence mechanisms in the body, thereby causing hypersusceptibility to certain exposures.
Most research related to susceptibility has dealt with genetic predisposition. Other factors play a role as well and have been partly neglected. For example, individuals with a chronic disease may be more sensitive to an occupational exposure. Also, if a disease process or previous exposure to toxic chemicals has caused some subclinical organ damage, then the capacity to withstand a new toxic exposure is likely to be less. Biochemical indicators of organ function may in this case be used as susceptibility biomarkers. Perhaps the best example regarding hypersusceptibility relates to allergic responses. If an individual has become sensitized to a particular exposure, then specific antibodies can be detected in serum. Even if the individual has not become sensitized, other current or past exposures may add to the risk of developing an adverse effect related to an occupational exposure.
A major problem is to determine the joint effect of mixed exposures at work. In addition, personal habits and drug use may result in an increased susceptibility. For example, tobacco smoke usually contains a considerable amount of cadmium. Thus, with occupational exposure to cadmium, a heavy smoker who has accumulated substantial amounts of this metal in the body will be at increased risk of developing cadmium-related kidney disease.
Application in Occupational Health
Biomarkers are extremely useful in toxicological research, and many may be applicable in biological monitoring. Nonetheless, the limitations must also be recognized. Many biomarkers have so far been studied only in laboratory animals. Toxicokinetic patterns in other species may not necessarily reflect the situation in human beings, and extrapolation may require confirmatory studies in human volunteers. Also, account must be taken of individual variations due to genetic or constitutional factors.
In some cases, exposure biomarkers may not at all be feasible (e.g., for chemicals which are short-lived in vivo). Other chemicals may be stored in, or may affect, organs which cannot be accessed by routine procedures, such as the nervous system. The route of exposure may also affect the distribution pattern and therefore also the biomarker measurement and its interpretation. For example, direct exposure of the brain via the olfactory nerve is likely to escape detection by measurement of exposure biomarkers. As to effect biomarkers, many of them are not at all specific, and the change can be due to a variety of causes, including lifestyle factors. Perhaps in particular with the susceptibility biomarkers, interpretation must be very cautious at the moment, as many uncertainties remain about the overall health significance of individual genotypes.
In occupational health, the ideal biomarker should satisfy several requirements. First of all, sample collection and analysis must be simple and reliable. For optimal analytical quality, standardization is needed, but the specific requirements vary considerably. Major areas of concern include: preparation of the in- dividual, sampling procedure and sample handling, and measurement procedure; the latter encompasses technical factors, such as calibration and quality assurance procedures, and individual- related factors, such as education and training of operators.
For documentation of analytical validity and traceability, reference materials should be based on relevant matrices and with appropriate concentrations of toxic substances or relevant metabolites at appropriate levels. For biomarkers to be used for biological monitoring or for diagnostic purposes, the responsible laboratories must have well-documented analytical procedures with defined performance characteristics, and accessible records to allow verification of the results. At the same time, nonetheless, the economics of characterizing and using reference materials to supplement quality assurance procedures in general must be considered. Thus, the achievable quality of results, and the uses to which they are put, have to be balanced against the added costs of quality assurance, including reference materials, manpower and instrumentation.
Another requirement is that the biomarker should be specific, at least under the circumstances of the study, for a particular type of exposure, with a clear-cut relationship to the degree of exposure. Otherwise, the result of the biomarker measurement may be too difficult to interpret. For proper interpretation of the measurement result of an exposure biomarker, the diagnostic validity must be known (i.e., the translation of the biomarker value into the magnitude of possible health risks). In this area, metals serve as a paradigm for biomarker research. Recent research has demonstrated the complexity and subtlety of dose-response relationships, with considerable difficulty in identifying no-effect levels and therefore also in defining tolerable exposures. However, this kind of research has also illustrated the types of investigation and the refinement that are necessary to uncover the relevant information. For most organic compounds, quantitative associations between exposures and the corresponding adverse health effects are not yet available; in many cases, even the primary target organs are not known for sure. In addition, evaluation of toxicity data and biomarker concentrations is often complicated by exposure to mixtures of substances, rather than exposure to a single compound at the time.
Before the biomarker is applied for occupational health purposes, some additional considerations are necessary. First, the biomarker must reflect a subclinical and reversible change only. Second, given that the biomarker results can be interpreted with regard to health risks, then preventive efforts should be available and should be considered realistic in case the biomarker data suggests a need to reduce the exposure. Third, the practical use of the biomarker must be generally regarded as ethically acceptable.
Industrial hygiene measurements may be compared with applicable exposure limits. Likewise, results on exposure biomarkers or effect biomarkers may be compared to biological action limits, sometimes referred to as biological exposure indices. Such limits should be based on the best advice of clinicians and scientists from appropriate disciplines, and responsible administrators as “risk managers” should then take into account relevant ethical, social, cultural and economic factors. The scientific basis should, if possible, include dose-response relationships supplemented by information on variations in susceptibility within the population at risk. In some countries, workers and members of the general public are involved in the standard-setting process and provide important input, particularly when scientific uncertainty is considerable. One of the major uncertainties is how to define an adverse health effect that should be prevented—for example, whether adduct formation as an exposure biomarker by itself represents an adverse effect (i.e., effect biomarker) that should be prevented. Difficult questions are likely to arise when deciding whether it is ethically defensible, for the same compound, to have different limits for adventitious exposure, on the one hand, and occupational exposure, on the other.
The information generated by the use of biomarkers should generally be conveyed to the individuals examined within the physician-patient relationship. Ethical concerns must in particular be considered in connection with highly experimental biomarker analyses that cannot currently be interpreted in detail in terms of actual health risks. For the general population, for example, limited guidance exists at present with regard to interpretation of exposure biomarkers other than the blood-lead concentration. Also of importance is the confidence in the data generated (i.e., whether appropriate sampling has been done, and whether sound quality assurance procedures have been utilized in the laboratory involved). An additional area of special worry relates to individual hypersusceptibility. These issues must be taken into account when providing the feedback from the study.
All sectors of society affected by, or concerned with carrying out, a biomarker study need to be involved in the decision-making process on how to handle the information generated by the study. Specific procedures to prevent or overcome inevitable ethical conflicts should be developed within the legal and social frameworks of the region or country. However, each situation represents a different set of questions and pitfalls, and no single procedure for public involvement can be developed to cover all applications of exposure biomarkers.
Genetic toxicity assessment is the evaluation of agents for their ability to induce any of three general types of changes (mutations) in the genetic material (DNA): gene, chromosomal and genomic. In organisms such as humans, the genes are composed of DNA, which consists of individual units called nucleotide bases. The genes are arranged in discrete physical structures called chromosomes. Genotoxicity can result in significant and irreversible effects upon human health. Genotoxic damage is a critical step in the induction of cancer and it can also be involved in the induction of birth defects and foetal death. The three classes of mutations mentioned above can occur within either of the two types of tissues possessed by organisms such as humans: sperm or eggs (germ cells) and the remaining tissue (somatic cells).
Assays that measure gene mutation are those that detect the substitution, addition or deletion of nucleotides within a gene. Assays that measure chromosomal mutation are those that detect breaks or chromosomal rearrangements involving one or more chromosomes. Assays that measure genomic mutation are those that detect changes in the number of chromosomes, a condition called aneuploidy. Genetic toxicity assessment has changed considerably since the development by Herman Muller in 1927 of the first assay to detect genotoxic (mutagenic) agents. Since then, more than 200 assays have been developed that measure mutations in DNA; however, fewer than ten assays are used commonly today for genetic toxicity assessment. This article reviews these assays, describes what they measure, and explores the role of these assays in toxicity assessment.
Identification of Cancer HazardsPrior to the Development of the Fieldof Genetic Toxicology
Genetic toxicology has become an integral part of the overall risk assessment process and has gained in stature in recent times as a reliable predictor for carcinogenic activity. However, prior to the development of genetic toxicology (before 1970), other methods were and are still being used to identify potential cancer hazards to humans. There are six major categories of methods currently used for identifying human cancer risks: epidemiological studies, long-term in vivo bioassays, mid-term in vivo bioassays, short-term in vivo and in vitro bioassays, artificial intelligence (structure-activity), and mechanism-based inference.
Table 1. Advantages and disadvantages of current methods for identifying human cancer risks
|Epidemiological studies||(1) humans are ultimate indicators of disease;
(2) evaluate sensitive or susceptible populations;
(3) occupational exposure cohorts; (4) environmental sentinel alerts
|(1) generally retrospective (death certificates, recall biases, etc.); (2) insensitive, costly, lengthy; (3) reliable exposure data sometimes unavailable or difficult to obtain; (4) combined, multiple and complex exposures; lack of appropriate control cohorts; (5) experiments on humans not done; (6) cancer detection, not prevention|
|Long-term in vivo bioassays||(1) prospective and retrospective (validation) evaluations; (2) excellent correlation with identified human carcinogens; (3) exposure levels and conditions known; (4) identifies chemical toxicity and carcinogenicity effects; (5) results obtained relatively quickly; (6) qualitative comparisons among chemical classes; (7) integrative and interactive biologic systems related closely to humans||(1) rarely replicated, resource intensive; (3) limited facilities suitable for such experiments; (4) species extrapolation debate; (5) exposures used are often at levels far in excess of those experienced by humans; (6) single-chemical exposure does not mimic human exposures, which are generally to multiple chemicals simultaneously|
|Mid- and short-term in vivo and in vitro bioassays||(1) more rapid and less expensive than other assays; (2) large samples that are easily replicated;
(3) biologically meaningful end points are measured (mutation, etc.); (4) can be used as screening assays to select chemicals for long-term bioassays
|(1) in vitro not fully predictive of in vivo; (2) usually organism or organ specific; (3) potencies not comparable to whole animals or humans|
|Chemical structure–biological activity associations||(1) relatively easy, rapid, and inexpensive; (2) reliable for certain chemical classes (e.g., nitrosamines and benzidine dyes); (3) developed from biological data but not dependent on additional biological experimentation||(1) not “biological”; (2) many exceptions to formulated rules; (3) retrospective and rarely (but becoming) prospective|
|Mechanism-based inferences||(1) reasonably accurate for certain classes of chemicals; (2) permits refinements of hypotheses; (3) can orient risk assessments to sensitive populations||(1) mechanisms of chemical carcinogenesis undefined, multiple, and likely chemical or class specific; (2) may fail to highlight exceptions to general mechanisms|
Rationale and Conceptual Basisfor Genetic Toxicology Assays
Although the exact types and numbers of assays used for genetic toxicity assessment are constantly evolving and vary from country to country, the most common ones include assays for (1) gene mutation in bacteria and/or cultured mammalian cells and (2) chromosomal mutation in cultured mammalian cells and/or bone marrow within living mice. Some of the assays within this second category can also detect aneuploidy. Although these assays do not detect mutations in germ cells, they are used primarily because of the extra cost and complexity of performing germ-cell assays. Nonetheless, germ-cell assays in mice are used when information about germ-cell effects is desired.
Systematic studies over a 25-year period (1970-1995), especially at the US National Toxicology Program in North Carolina, have resulted in the use of a discrete number of assays for detecting the mutagenic activity of agents. The rationale for evaluating the usefulness of the assays was based on their ability to detect agents that cause cancer in rodents and that are suspected of causing cancer in humans (i.e., carcinogens). This is because studies during the past several decades have indicated that cancer cells contain mutations in certain genes and that many carcinogens are also mutagens. Thus, cancer cells are viewed as containing somatic-cell mutations, and carcinogenesis is viewed as a type of somatic-cell mutagenesis.
The genetic toxicity assays used most commonly today have been selected not only because of their large database, relatively low cost, and ease of performance, but because they have been shown to detect many rodent and, presumptively, human carcinogens. Consequently, genetic toxicity assays are used to predict the potential carcinogenicity of agents.
An important conceptual and practical development in the field of genetic toxicology was the recognition that many carcinogens were modified by enzymes within the body, creating altered forms (metabolites) that were frequently the ultimate carcinogenic and mutagenic form of the parent chemical. To duplicate this metabolism in a petri dish, Heinrich Malling showed that the inclusion of a preparation from rodent liver contained many of the enzymes necessary to perform this metabolic conversion or activation. Thus, many genetic toxicity assays performed in dishes or tubes (in vitro) employ the addition of similar enzyme preparations. Simple preparations are called S9 mix, and purified preparations are called microsomes. Some bacterial and mammalian cells have now been genetically engineered to contain some of the genes from rodents or humans that produce these enzymes, reducing the need to add S9 mix or microsomes.
Genetic Toxicology Assays and Techniques
The primary bacterial systems used for genetic toxicity screening are the Salmonella (Ames) mutagenicity assay and, to a much lesser extent, strain WP2 of Escherichia coli. Studies in the mid-1980s indicated that the use of only two strains of the Salmonella system (TA98 and TA100) were sufficient to detect approximately 90% of the known Salmonella mutagens. Thus, these two strains are used for most screening purposes; however, various other strains are available for more extensive testing.
These assays are performed in a variety of ways, but two general procedures are the plate-incorporation and liquid-suspension assays. In the plate-incorporation assay, the cells, the test chemical and (when desired) the S9 are added together into a liquefied agar and poured onto the surface of an agar petri plate. The top agar hardens within a few minutes, and the plates are incubated for two to three days, after which time mutant cells have grown to form visually detectable clusters of cells called colonies, which are then counted. The agar medium contains selective agents or is composed of ingredients such that only the newly mutated cells will grow. The liquid-incubation assay is similar, except the cells, test agent, and S9 are incubated together in liquid that does not contain liquefied agar, and then the cells are washed free of the test agent and S9 and seeded onto the agar.
Mutations in cultured mammalian cells are detected primarily in one of two genes: hprt and tk. Similar to the bacterial assays, mammalian cell lines (developed from rodent or human cells) are exposed to the test agent in plastic culture dishes or tubes and then are seeded into culture dishes that contain medium with a selective agent that permits only mutant cells to grow. The assays used for this purpose include the CHO/HPRT, the TK6, and the mouse lymphoma L5178Y/TK+/- assays. Other cell lines containing various DNA repair mutations as well as containing some human genes involved in metabolism are also used. These systems permit the recovery of mutations within the gene (gene mutation) as well as mutations involving regions of the chromosome flanking the gene (chromosomal mutation). However, this latter type of mutation is recovered to a much greater extent by the tk gene systems than by the hprt gene systems due to the location of the tk gene.
Similar to the liquid-incubation assay for bacterial mutagenicity, mammalian cell mutagenicity assays generally involve the exposure of the cells in culture dishes or tubes in the presence of the test agent and S9 for several hours. The cells are then washed, cultured for several more days to allow the normal (wild-type) gene products to be degraded and the newly mutant gene products to be expressed and accumulate, and then they are seeded into medium containing a selective agent that permits only the mutant cells to grow. Like the bacterial assays, the mutant cells grow into visually detectable colonies that are then counted.
Chromosomal mutation is identified primarily by cytogenetic assays, which involve exposing rodents and/or rodent or human cells in culture dishes to a test chemical, allowing one or more cell divisions to occur, staining the chromosomes, and then visually examining the chromosomes through a microscope to detect alterations in the structure or number of chromosomes. Although a variety of endpoints can be examined, the two that are currently accepted by regulatory agencies as being the most meaningful are chromosomal aberrations and a subcategory called micronuclei.
Considerable training and expertise are required to score cells for the presence of chromosomal aberrations, making this a costly procedure in terms of time and money. In contrast, micronuclei require little training, and their detection can be automated. Micronuclei appear as small dots within the cell that are distinct from the nucleus, which contains the chromosomes. Micronuclei result from either chromosome breakage or from aneuploidy. Because of the ease of scoring micronuclei compared to chromosomal aberrations, and because recent studies indicate that agents that induce chromosomal aberrations in the bone marrow of living mice generally induce micronuclei in this tissue, micronuclei are now commonly measured as an indication of the ability of an agent to induce chromosomal mutation.
Although germ-cell assays are used far less frequently than the other assays described above, they are indispensable in determining whether an agent poses a risk to the germ cells, mutations in which can lead to health effects in succeeding generations. The most commonly used germ-cell assays are in mice, and involve systems that detect (1) heritable translocations (exchanges) among chromosomes (heritable translocation assay), (2) gene or chromosomal mutations involving specific genes (visible or biochemical specific-locus assays), and (3) mutations that affect viability (dominant lethal assay). As with the somatic-cell assays, the working assumption with the germ-cell assays is that agents positive in these assays are presumed to be potential human germ-cell mutagens.
Current Status and Future Prospects
Recent studies have indicated that only three pieces of information were necessary to detect approximately 90% of a set of 41 rodent carcinogens (i.e., presumptive human carcinogens and somatic-cell mutagens). These included (1) knowledge of the chemical structure of agent, especially if it contains electrophilic moieties (see section on structure-activity relationships); (2) Salmonella mutagenicity data; and (3) data from a 90-day chronic toxicity assay in rodents (mice and rats). Indeed, essentially all of the IARC-declared human carcinogens are detectable as mutagens using just the Salmonella assay and the mouse-bone marrow micronucleus assay. The use of these mutagenicity assays for detecting potential human carcinogens is supported further by the finding that most human carcinogens are carcinogenic in both rats and mice (trans-species carcinogens) and that most trans- species carcinogens are mutagenic in Salmonella and/or induce micronuclei in mouse bone marrow.
With advances in DNA technology, the human genome project, and an improved understanding of the role of mutation in cancer, new genotoxicity assays are being developed that will likely be incorporated into standard screening procedures. Among these are the use of transgenic cells and rodents. Transgenic systems are those in which a gene from another species has been introduced into a cell or organism. For example, transgenic mice are now in experimental use that permit the detection of mutation in any organ or tissue of the animal, based on the introduction of a bacterial gene into the mouse. Bacterial cells, such as Salmonella, and mammalian cells (including human cell lines) are now available that contain genes involved in the metabolism of carcinogenic/mutagenic agents, such as the P450 genes. Molecular analysis of the actual mutations induced in the trans-gene within transgenic rodents, or within native genes such as hprt, or the target genes within Salmonella can now be performed, so that the exact nature of the mutations induced by the chemicals can be determined, providing insights into the mechanism of action of the chemical and allowing comparisons to mutations in humans presumptively exposed to the agent.
Molecular advances in cytogenetics now permit more detailed evaluation of chromosomal mutations. These include the use of probes (small pieces of DNA) that attach (hybridize) to specific genes. Rearrangements of genes on the chromosome can then be revealed by the altered location of the probes, which are fluorescent and easily visualized as colored sectors on the chromosomes. The single-cell gel electrophoresis assay for DNA breakage (commonly called the “comet” assay) permits the detection of DNA breaks within single cells and may become an extremely useful tool in combination with cytogenetic techniques for detecting chromosomal damage.
After many years of use and the generation of a large and systematically developed database, genetic toxicity assessment can now be done with just a few assays for relatively small cost in a short period of time (a few weeks). The data produced can be used to predict the ability of an agent to be a rodent and, presumptively, human carcinogen/somatic-cell mutagen. Such an ability makes it possible to limit the introduction into the environment of mutagenic and carcinogenic agents and to develop alternative, nonmutagenic agents. Future studies should lead to even better methods with greater predictivity than the current assays.
The emergence of sophisticated technologies in molecular and cellular biology has spurred a relatively rapid evolution in the life sciences, including toxicology. In effect, the focus of toxicology is shifting from whole animals and populations of whole animals to the cells and molecules of individual animals and humans. Since the mid-1980s, toxicologists have begun to employ these new methodologies in assessing the effects of chemicals on living systems. As a logical progression, such methods are being adapted for the purposes of toxicity testing. These scientific advances have worked together with social and economic factors to effect change in the evaluation of product safety and potential risk.
Economic factors are specifically related to the volume of materials that must be tested. A plethora of new cosmetics, pharmaceuticals, pesticides, chemicals and household products is introduced into the market every year. All of these products must be evaluated for their potential toxicity. In addition, there is a backlog of chemicals already in use that have not been adequately tested. The enormous task of obtaining detailed safety information on all of these chemicals using traditional whole animal testing methods would be costly in terms of both money and time, if it could even be accomplished.
There are also societal issues that relate to public health and safety, as well as increasing public concern about the use of animals for product safety testing. With regard to human safety, public interest and environmental advocacy groups have placed significant pressure on government agencies to apply more stringent regulations on chemicals. A recent example of this has been a movement by some environmental groups to ban chlorine and chlorine-containing compounds in the United States. One of the motivations for such an extreme action lies in the fact that most of these compounds have never been adequately tested. From a toxicological perspective, the concept of banning a whole class of diverse chemicals based simply on the presence of chlorine is both scientifically unsound and irresponsible. Yet, it is understandable that from the public’s perspective, there must be some assurance that chemicals released into the environment do not pose a significant health risk. Such a situation underscores the need for more efficient and rapid methods to assess toxicity.
The other societal concern that has impacted the area of toxicity testing is animal welfare. The growing number of animal protection groups throughout the world have voiced considerable opposition to the use of whole animals for product safety testing. Active campaigns have been waged against manufacturers of cosmetics, household and personal care products and pharmaceuticals in attempts to stop animal testing. Such efforts in Europe have resulted in the passage of the Sixth Amendment to Directive 76/768/EEC (the Cosmetics Directive). The consequence of this Directive is that cosmetic products or cosmetic ingredients that have been tested in animals after January 1, 1998 cannot be marketed in the European Union, unless alternative methods are insufficiently validated. While this Directive has no jurisdiction over the sale of such products in the United States or other countries, it will significantly affect those companies that have international markets that include Europe.
The concept of alternatives, which forms the basis for the development of tests other than those on whole animals, is defined by the three Rs: reduction in the numbers of animals used; refinement of protocols so that animals experience less stress or discomfort; and replacement of current animal tests with in vitro tests (i.e., tests done outside of the living animal), computer models or test on lower vertebrate or invertebrate species. The three Rs were introduced in a book published in 1959 by two British scientists, W.M.S. Russell and Rex Burch, The Principles of Humane Experimental Technique. Russell and Burch maintained that the only way in which valid scientific results could be obtained is through the humane treatment of animals, and believed that methods should be developed to reduce animal use and ultimately replace it. Interestingly, the principles outlined by Russell and Burch received little attention until the resurgence of the animal welfare movement in the mid-1970s. Today the concept of the three Rs is very much in the forefront with regard to research, testing and education.
In summary, the development of in vitro test methodologies has been influenced by a variety of factors that have converged over the last ten to 20 years. It is difficult to ascertain if any of these factors alone would have had such a profound effect on toxicity testing strategies.
Concept of In Vitro Toxicity Tests
This section will focus solely on in vitro methods for evaluating toxicity, as one of the alternatives to whole-animal testing. Additional non-animal alternatives such as computer modelling and quantitative structure-activity relationships are discussed in other articles of this chapter.
In vitro studies are generally conducted in animal or human cells or tissues outside of the body. In vitro literally means “in glass”, and refers to procedures carried out on living material or components of living material cultured in petri dishes or in test tubes under defined conditions. These may be contrasted with in vivo studies, or those carried out “in the living animal”. While it is difficult, if not impossible, to project the effects of a chemical on a complex organism when the observations are confined to a single type of cells in a dish, in vitro studies do provide a significant amount of information about intrinsic toxicity as well as cellular and molecular mechanisms of toxicity. In addition, they offer many advantages over in vivo studies in that they are generally less expensive and they may be conducted under more controlled conditions. Furthermore, despite the fact that small numbers of animals are still needed to obtain cells for in vitro cultures, these methods may be considered reduction alternatives (since many fewer animals are used compared to in vivo studies) and refinement alternatives (because they eliminate the need to subject the animals to the adverse toxic consequences imposed by in vivo experiments).
In order to interpret the results of in vitro toxicity tests, determine their potential usefulness in assessing toxicity and relate them to the overall toxicological process in vivo, it is necessary to understand which part of the toxicological process is being examined. The entire toxicological process consists of events that begin with the organism’s exposure to a physical or chemical agent, progress through cellular and molecular interactions and ultimately manifest themselves in the response of the whole organism. In vitro tests are generally limited to the part of the toxicological process that takes place at the cellular and molecular level. The types of information that may be obtained from in vitro studies include pathways of metabolism, interaction of active metabolites with cellular and molecular targets and potentially measurable toxic endpoints that can serve as molecular biomarkers for exposure. In an ideal situation, the mechanism of toxicity of each chemical from exposure to organismal manifestation would be known, such that the information obtained from in vitro tests could be fully interpreted and related to the response of the whole organism. However, this is virtually impossible, since relatively few complete toxicological mechanisms have been elucidated. Thus, toxicologists are faced with a situation in which the results of an in vitro test cannot be used as an entirely accurate prediction of in vivo toxicity because the mechanism is unknown. However, frequently during the process of developing an in vitro test, components of the cellular and molecular mechanism(s) of toxicity are elucidated.
One of the key unresolved issues surrounding the development and implementation of in vitro tests is related to the following consideration: should they be mechanistically based or is it sufficient for them to be descriptive? It is inarguably better from a scientific perspective to utilize only mechanistically based tests as replacements for in vivo tests. However in the absence of complete mechanistic knowledge, the prospect of developing in vitro tests to completely replace whole animal tests in the near future is almost nil. This does not, however, rule out the use of more descriptive types of assays as early screening tools, which is the case presently. These screens have resulted in a significant reduction in animal use. Therefore, until such time as more mechanistic information is generated, it may be necessary to employ to a more limited extent, tests whose results simply correlate well with those obtained in vivo.
In Vitro Tests for Cytotoxicity
In this section, several in vitro tests that have been developed to assess a chemical’s cytotoxic potential will be described. For the most part, these tests are easy to perform and analysis can be automated. One commonly used in vitro test for cytotoxicity is the neutral red assay. This assay is done on cells in culture, and for most applications, the cells can be maintained in culture dishes that contain 96 small wells, each 6.4mm in diameter. Since each well can be used for a single determination, this arrangement can accommodate multiple concentrations of the test chemical as well as positive and negative controls with a sufficient number of replicates for each. Following treatment of the cells with various concentrations of the test chemical ranging over at least two orders of magnitude (e.g., from 0.01mM to 1mM), as well as positive and negative control chemicals, the cells are rinsed and treated with neutral red, a dye that can be taken up and retained only by live cells. The dye may be added upon removal of the test chemical to determine immediate effects, or it may be added at various times after the test chemical is removed to determine cumulative or delayed effects. The intensity of the colour in each well corresponds to the number of live cells in that well. The colour intensity is measured by a spectrophotometer which may be equipped with a plate reader. The plate reader is programmed to provide individual measurements for each of the 96 wells of the culture dish. This automated methodology permits the investigator to rapidly perform a concentration-response experiment and to obtain statistically useful data.
Another relatively simple assay for cytotoxicity is the MTT test. MTT (3[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) is a tetrazolium dye that is reduced by mitochondrial enzymes to a blue colour. Only cells with viable mitochondria will retain the ability to carry out this reaction; therefore the colour intensity is directly related to the degree of mitochondrial integrity. This is a useful test to detect general cytotoxic compounds as well as those agents that specifically target mitochondria.
The measurement of lactate dehydrogenase (LDH) activity is also used as a broad-based assay for cytotoxicity. This enzyme is normally present in the cytoplasm of living cells and is released into the cell culture medium through leaky cell membranes of dead or dying cells that have been adversely affected by a toxic agent. Small amounts of culture medium may be removed at various times after chemical treatment of the cells to measure the amount of LDH released and determine a time course of toxicity. While the LDH release assay is a very general assessment of cytotoxicity, it is useful because it is easy to perform and it may be done in real time.
There are many new methods being developed to detect cellular damage. More sophisticated methods employ fluorescent probes to measure a variety of intracellular parameters, such as calcium release and changes in pH and membrane potential. In general, these probes are very sensitive and may detect more subtle cellular changes, thus reducing the need to use cell death as an endpoint. In addition, many of these fluorescent assays may be automated by the use of 96-well plates and fluorescent plate readers.
Once data have been collected on a series of chemicals using one of these tests, the relative toxicities may be determined. The relative toxicity of a chemical, as determined in an in vitro test, may be expressed as the concentration that exerts a 50% effect on the endpoint response of untreated cells. This determination is referred to as the EC50 (Effective Concentration for 50% of the cells) and may be used to compare toxicities of different chemicals in vitro. (A similar term used in evaluating relative toxicity is IC50, indicating the concentration of a chemical that causes a 50% inhibition of a cellular process, e.g., the ability to take up neutral red.) It is not easy to assess whether the relative in vitro toxicity of the chemicals is comparable to their relative in vivo toxicities, since there are so many confounding factors in the in vivo system, such as toxicokinetics, metabolism, repair and defence mechanisms. In addition, since most of these assays measure general cytotoxicity endpoints, they are not mechanistically based. Therefore, agreement between in vitro and in vivo relative toxicities is simply correlative. Despite the numerous complexities and difficulties in extrapolating from in vitro to in vivo, these in vitro tests are proving to be very valuable because they are simple and inexpensive to perform and may be used as screens to flag highly toxic drugs or chemicals at early stages of development.
Target Organ Toxicity
In vitro tests can also be used to assess specific target organ toxicity. There are a number of difficulties associated with designing such tests, the most notable being the inability of in vitro systems to maintain many of the features of the organ in vivo. Frequently, when cells are taken from animals and placed into culture, they tend either to degenerate quickly and/or to dedifferentiate, that is, lose their organ-like functions and become more generic. This presents a problem in that within a short period of time, usually a few days, the cultures are no longer useful for assessing organ-specific effects of a toxin.
Many of these problems are being overcome because of recent advances in molecular and cellular biology. Information that is obtained about the cellular environment in vivo may be utilized in modulating culture conditions in vitro. Since the mid-1980s, new growth factors and cytokines have been discovered, and many of these are now available commercially. Addition of these factors to cells in culture helps to preserve their integrity and may also help to retain more differentiated functions for longer periods of time. Other basic studies have increased the knowledge of the nutritional and hormonal requirements of cells in culture, so that new media may be formulated. Recent advances have also been made in identifying both naturally occurring and artificial extracellular matrices on which cells may be cultured. Culture of cells on these different matrices can have profound effects on both their structure and function. A major advantage derived from this knowledge is the ability to intricately control the environment of cells in culture and individually examine the effects of these factors on basic cell processes and on their responses to different chemical agents. In short, these systems can provide great insight into organ-specific mechanisms of toxicity.
Many target organ toxicity studies are conducted in primary cells, which by definition are freshly isolated from an organ, and usually exhibit a finite lifetime in culture. There are many advantages to having primary cultures of a single cell type from an organ for toxicity assessment. From a mechanistic perspective, such cultures are useful for studying specific cellular targets of a chemical. In some instances, two or more cell types from an organ may be cultured together, and this provides an added advantage of being able to look at cell-cell interactions in response to a toxin. Some co-culture systems for skin have been engineered so that they form a three dimensional structure resembling skin in vivo. It is also possible to co-culture cells from different organs—for example, liver and kidney. This type of culture would be useful in assessing the effects specific to kidney cells, of a chemical that must be bioactivated in the liver.
Molecular biological tools have also played an important role in the development of continuous cell lines that can be useful for target organ toxicity testing. These cell lines are generated by transfecting DNA into primary cells. In the transfection procedure, the cells and the DNA are treated such that the DNA can be taken up by the cells. The DNA is usually from a virus and contains a gene or genes that, when expressed, allow the cells to become immortalized (i.e., able to live and grow for extended periods of time in culture). The DNA can also be engineered so that the immortalizing gene is controlled by an inducible promoter. The advantage of this type of construct is that the cells will divide only when they receive the appropriate chemical stimulus to allow expression of the immortalizing gene. An example of such a construct is the large T antigen gene from Simian Virus 40 (SV40) (the immortalizing gene), preceded by the promoter region of the metallothionein gene, which is induced by the presence of a metal in the culture medium. Thus, after the gene is transfected into the cells, the cells may be treated with low concentrations of zinc to stimulate the MT promoter and turn on the expression of the T antigen gene. Under these conditions, the cells proliferate. When zinc is removed from the medium, the cells stop dividing and under ideal conditions return to a state where they express their tissue-specific functions.
The ability to generate immortalized cells combined with the advances in cell culture technology have greatly contributed to the creation of cell lines from many different organs, including brain, kidney and liver. However, before these cell lines may be used as a surrogate for the bona fide cell types, they must be carefully characterized to determine how “normal” they really are.
Other in vitro systems for studying target organ toxicity involve increasing complexity. As in vitro systems progress in complexity from single cell to whole organ culture, they become more comparable to the in vivo milieu, but at the same time they become much more difficult to control given the increased number of variables. Therefore, what may be gained in moving to a higher level of organization can be lost in the inability of the researcher to control the experimental environment. Table 1 compares some of the characteristics of various in vitro systems that have been used to study hepatotoxicity.
Table 1. Comparison of in vitro systems for hepatotoxicity studies
(level of interaction)
|Ability to retain liver-specific functions||Potential duration of culture||Ability to control environment|
|Immortalized cell lines||some cell to cell (varies with cell line)||poor to good (varies with cell line)||indefinite||excellent|
|Primary hepatocyte cultures||cell to cell||fair to excellent (varies with culture conditions)||days to weeks||excellent|
|Liver cell co-cultures||cell to cell (between the same and different cell types)||good to excellent||weeks||excellent|
|Liver slices||cell to cell (among all cell types)||good to excellent||hours to days||good|
|Isolated, perfused liver||cell to cell (among all cell types), and intra-organ||excellent||hours||fair|
Precision-cut tissue slices are being used more extensively for toxicological studies. There are new instruments available that enable the researcher to cut uniform tissue slices in a sterile environment. Tissue slices offer some advantage over cell culture systems in that all of the cell types of the organ are present and they maintain their in vivo architecture and intercellular communication. Thus, in vitro studies may be conducted to determine the target cell type within an organ as well as to investigate specific target organ toxicity. A disadvantage of the slices is that they degenerate rapidly after the first 24 hours of culture, mainly due to poor diffusion of oxygen to the cells on the interior of the slices. However, recent studies have indicated that more efficient aeration may be achieved by gentle rotation. This, together with the use of a more complex medium, allows the slices to survive for up to 96 hours.
Tissue explants are similar in concept to tissue slices and may also be used to determine the toxicity of chemicals in specific target organs. Tissue explants are established by removing a small piece of tissue (for teratogenicity studies, an intact embryo) and placing it into culture for further study. Explant cultures have been useful for short-term toxicity studies including irritation and corrosivity in skin, asbestos studies in trachea and neurotoxicity studies in brain tissue.
Isolated perfused organs may also be used to assess target organ toxicity. These systems offer an advantage similar to that of tissue slices and explants in that all cell types are present, but without the stress to the tissue introduced by the manipulations involved in preparing slices. In addition, they allow for the maintenance of intra-organ interactions. A major disadvantage is their short-term viability, which limits their use for in vitro toxicity testing. In terms of serving as an alternative, these cultures may be considered a refinement since the animals do not experience the adverse consequences of in vivo treatment with toxicants. However, their use does not significantly decrease the numbers of animals required.
In summary, there are several types of in vitro systems available for assessing target organ toxicity. It is possible to acquire much information about mechanisms of toxicity using one or more of these techniques. The difficulty remains in knowing how to extrapolate from an in vitro system, which represents a relatively small part of the toxicological process, to the whole process occurring in vivo.
In Vitro Tests for Ocular Irritation
Perhaps the most contentious whole-animal toxicity test from an animal welfare perspective is the Draize test for eye irritation, which is conducted in rabbits. In this test, a small fixed dose of a chemical is placed in one of the rabbit’s eyes while the other eye is used as a control. The degree of irritation and inflammation is scored at various times after exposure. A major effort is being made to develop methodologies to replace this test, which has been criticized not only for humane reasons, but also because of the subjectivity of the observations and variability of the results. It is interesting to note that despite the harsh criticism the Draize test has received, it has proven to be remarkably successful in predicting human eye irritants, particularly slightly to moderately irritating substances, that are difficult to identify by other methods. Thus, the demands on in vitro alternatives are great.
The quest for alternatives to the Draize test is a complicated one, albeit one that is predicted to be successful. Numerous in vitro and other alternatives have been developed and in some cases they have been implemented. Refinement alternatives to the Draize test, which by definition, are less painful or distressful to the animals, include the Low Volume Eye Test, in which smaller amounts of test materials are placed in the rabbits’ eyes, not only for humane reasons, but to more closely mimic the amounts to which people may actually be accidentally exposed. Another refinement is that substances which have a pH less than 2 or greater than 11.5 are no longer tested in animals since they are known to be severely irritating to the eye.
Between 1980 and 1989, there has been an estimated 87% decline in the number of rabbits used for eye irritation testing of cosmetics. In vitro tests have been incorporated as part of a tier-testing approach to bring about this vast reduction in whole-animal tests. This approach is a multi-step process that begins with a thorough examination of the historical eye irritation data and physical and chemical analysis of the chemical to be evaluated. If these two processes do not yield enough information, then a battery of in vitro tests is performed. The additional data obtained from the in vitro tests might then be sufficient to assess the safety of the substance. If not, then the final step would be to perform limited in vivo tests. It is easy to see how this approach can eliminate or at least drastically reduce the numbers of animals needed to predict the safety of a test substance.
The battery of in vitro tests that is used as part of this tier-testing strategy depends upon the needs of the particular industry. Eye irritation testing is done by a wide variety of industries from cosmetics to pharmaceuticals to industrial chemicals. The type of information required by each industry varies and therefore it is not possible to define a single battery of in vitro tests. A test battery is generally designed to assess five parameters: cytotoxicity, changes in tissue physiology and biochemistry, quantitative structure-activity relationships, inflammation mediators, and recovery and repair. An example of a test for cytotoxicity, which is one possible cause for irritation, is the neutral red assay using cultured cells (see above). Changes in cellular physiology and biochemistry resulting from exposure to a chemical may be assayed in cultures of human corneal epithelial cells. Alternatively, investigators have also used intact or dissected bovine or chicken eyeballs obtained from slaughterhouses. Many of the endpoints measured in these whole organ cultures are the same as those measured in vivo, such as corneal opacity and corneal swelling.
Inflammation is frequently a component of chemical-induced eye injury, and there are a number of assays available to examine this parameter. Various biochemical assays detect the presence of mediators released during the inflammatory process such as arachidonic acid and cytokines. The chorioallantoic membrane (CAM) of the hen’s egg may also be used as an indicator of inflammation. In the CAM assay, a small piece of the shell of a ten-to-14-day chick embryo is removed to expose the CAM. The chemical is then applied to the CAM and signs of inflammation, such as vascular hemorrhaging, are scored at various times thereafter.
One of the most difficult in vivo processes to assess in vitro is recovery and repair of ocular injury. A newly developed instrument, the silicon microphysiometer, measures small changes in extracellular pH and can been used to monitor cultured cells in real time. This analysis has been shown to correlate fairly well with in vivo recovery and has been used as an in vitro test for this process. This has been a brief overview of the types of tests being employed as alternatives to the Draize test for ocular irritation. It is likely that within the next several years a complete series of in vitro test batteries will be defined and each will be validated for its specific purpose.
The key to regulatory acceptance and implementation of in vitro test methodologies is validation, the process by which the credibility of a candidate test is established for a specific purpose. Efforts to define and coordinate the validation process have been made both in the United States and in Europe. The European Union established the European Centre for the Validation of Alternative Methods (ECVAM) in 1993 to coordinate efforts there and to interact with American organizations such as the Johns Hopkins Centre for Alternatives to Animal Testing (CAAT), an academic centre in the United States, and the Interagency Coordinating Committee for the Validation of Alternative Methods (ICCVAM), composed of representatives from the National Institutes of Health, the US Environmental Protection Agency, the US Food and Drug Administration and the Consumer Products Safety Commission.
Validation of in vitro tests requires substantial organization and planning. There must be consensus among government regulators and industrial and academic scientists on acceptable procedures, and sufficient oversight by a scientific advisory board to ensure that the protocols meet set standards. The validation studies should be performed in a series of reference laboratories using calibrated sets of chemicals from a chemical bank and cells or tissues from a single source. Both intralaboratory repeatability and interlaboratory reproducibility of a candidate test must be demonstrated and the results subjected to appropriate statistical analysis. Once the results from the different components of the validation studies have been compiled, the scientific advisory board can make recommendations on the validity of the candidate test(s) for a specific purpose. In addition, results of the studies should be published in peer-reviewed journals and placed in a database.
The definition of the validation process is currently a work in progress. Each new validation study will provide information useful to the design of the next study. International communication and cooperation are essential for the expeditious development of a widely acceptable series of protocols, particularly given the increased urgency imposed by the passage of the EC Cosmetics Directive. This legislation may indeed provide the needed impetus for a serious validation effort to be undertaken. It is only through completion of this process that the acceptance of in vitro methods by the various regulatory communities can commence.
This article has provided a broad overview of the current status of in vitro toxicity testing. The science of in vitro toxicology is relatively young, but it is growing exponentially. The challenge for the years ahead is to incorporate the mechanistic knowledge generated by cellular and molecular studies into the vast inventory of in vivo data to provide a more complete description of toxicological mechanisms as well as to establish a paradigm by which in vitro data may be used to predict toxicity in vivo. It will only be through the concerted efforts of toxicologists and government representatives that the inherent value of these in vitro methods can be realized.
Structure activity relationships (SAR) analysis is the utilization of information on the molecular structure of chemicals to predict important characteristics related to persistence, distribution, uptake and absorption, and toxicity. SAR is an alternative method of identifying potential hazardous chemicals, which holds promise of assisting industries and governments in prioritizing substances for further evaluation or for early-stage decision making for new chemicals. Toxicology is an increasingly expensive and resource-intensive undertaking. Increased concerns over the potential for chemicals to cause adverse effects in exposed human populations have prompted regulatory and health agencies to expand the range and sensitivity of tests to detect toxicological hazards. At the same time, the real and perceived burdens of regulation upon industry have provoked concerns for the practicality of toxicity testing methods and data analysis. At present, the determination of chemical carcinogenicity depends upon lifetime testing of at least two species, both sexes, at several doses, with careful histopathological analysis of multiple organs, as well as detection of preneoplastic changes in cells and target organs. In the United States, the cancer bioassay is estimated to cost in excess of $3 million (1995 dollars).
Even with unlimited financial resources, the burden of testing the approximately 70,000 existing chemicals produced in the world today would exceed the available resources of trained toxicologists. Centuries would be required to complete even a first tier evaluation of these chemicals (NRC 1984). In many countries ethical concerns over the use of animals in toxicity testing have increased, bringing additional pressures upon the uses of standard methods of toxicity testing. SAR has been widely used in the pharmaceutical industry to identify molecules with potential for beneficial use in treatment (Hansch and Zhang 1993). In environmental and occupational health policy, SAR is used to predict the dispersion of compounds in the physical-chemical environment and to screen new chemicals for further evaluation of potential toxicity. Under the US Toxic Substances Control Act (TSCA), the EPA has used since 1979 an SAR approach as a “first screen” of new chemicals in the premanufacture notification (PMN) process; Australia uses a similar approach as part of its new chemicals notification (NICNAS) procedure. In the US SAR analysis is an important basis for determining that there is a reasonable basis to conclude that manufacture, processing, distribution, use or disposal of the substance will present an unreasonable risk of injury to human health or the environment, as required by Section 5(f) of TSCA. On the basis of this finding, EPA can then require actual tests of the substance under Section 6 of TSCA.
Rationale for SAR
The scientific rationale for SAR is based upon the assumption that the molecular structure of a chemical will predict important aspects of its behaviour in physical-chemical and biological systems (Hansch and Leo 1979).
The SAR review process includes identification of the chemical structure, including empirical formulations as well as the pure compound; identification of structurally analogous substances; searching databases and literature for information on structural analogs; and analysis of toxicity and other data on structural analogs. In some rare cases, information on the structure of the compound alone can be sufficient to support some SAR analysis, based upon well-understood mechanisms of toxicity. Several databases on SAR have been compiled, as well as computer-based methods for molecular structure prediction.
With this information, the following endpoints can be estimated with SAR:
It should be noted that SAR methods do not exist for such important health endpoints as carcinogenicity, developmental toxicity, reproductive toxicity, neurotoxicity, immunotoxicity or other target organ effects. This is due to three factors: the lack of a large database upon which to test SAR hypotheses, lack of knowledge of structural determinants of toxic action, and the multiplicity of target cells and mechanisms that are involved in these endpoints (see “The United States approach to risk assessment of reproductive toxicants and neurotoxic agents”). Some limited attempts to utilize SAR for predicting pharmacokinetics using information on partition coefficients and solubility (Johanson and Naslund 1988). More extensive quantitative SAR has been done to predict P450-dependent metabolism of a range of compounds and binding of dioxin- and PCB-like molecules to the cytosolic “dioxin” receptor (Hansch and Zhang 1993).
SAR has been shown to have varying predictability for some of the endpoints listed above, as shown in table 1. This table presents data from two comparisons of predicted activity with actual results obtained by empirical measurement or toxicity testing. SAR as conducted by US EPA experts performed more poorly for predicting physical-chemical properties than for predicting biological activity, including biodegradation. For toxicity endpoints, SAR performed best for predicting mutagenicity. Ashby and Tennant (1991) in a more extended study also found good predictability of short-term genotoxicity in their analysis of NTP chemicals. These findings are not surprising, given current understanding of molecular mechanisms of genotoxicity (see “Genetic toxicology”) and the role of electrophilicity in DNA binding. In contrast, SAR tended to underpredict systemic and subchronic toxicity in mammals and to overpredict acute toxicity to aquatic organisms.
Table 1. Comparison of SAR and test data: OECD/NTP analyses
|Endpoint||Agreement (%)||Disagreement (%)||Number|
|Acute mammalian toxicity (LD50 )||80||201||142|
|Carcinogenicity3 : Two year bioassay||72–954||—||301|
Source: Data from OECD, personal communication C. Auer ,US EPA. Only those endpoints for which comparable SAR predictions and actual test data were available were used in this analysis. NTP data are from Ashby and Tennant 1991.
1 Of concern was the failure by SAR to predict acute toxicity in 12% of the chemicals tested.
2 OECD data, based on Ames test concordance with SAR
3 NTP data, based on genetox assays compared to SAR predictions for several classes of “structurally alerting chemicals”.
4 Concordance varies with class; highest concordance was with aromatic amino/nitro compounds; lowest with “miscellaneous” structures.
For other toxic endpoints, as noted above, SAR has less demonstrable utility. Mammalian toxicity predictions are complicated by the lack of SAR for toxicokinetics of complex molecules. Nevertheless, some attempts have been made to propose SAR principles for complex mammalian toxicity endpoints (for instance, see Bernstein (1984) for an SAR analysis of potential male reproductive toxicants). In most cases, the database is too small to permit rigorous testing of structure-based predictions.
At this point it may be concluded that SAR may be useful mainly for prioritizing the investment of toxicity testing resources or for raising early concerns about potential hazard. Only in the case of mutagenicity is it likely that SAR analysis by itself can be utilized with reliability to inform other decisions. For no endpoint is it likely that SAR can provide the type of quantitative information required for risk assessment purposes as discussed elsewhere in this chapter and Encyclopaedia.